연령 추정을 위한 주요 주름영역 선택에 관한 연구

In this paper, we proposed a new method to search the optional set of wrinkle areas, which maximize the accuracy while reducing the number of areas. We calculate a wrinkle density of each of 8 wrinkle areas and use them to estimate the age of a face. Support Vector Machine is used to predict the age of a face using wrinkle densities as features for classification. Backward elimination method is used to reduce the number of features. We also provided experiment result, in which only 4 percent of accuracy is decreased while using only 2 features instead of 8.